Hi Timothy, On Saturday 29 January 2005 11:22, you wrote: > Microsoft specifies a subset of VGA that must be supported for boot > and stuff like that. I don't know what that specification is, but > that is the subset we should implement so that Windows can display > what it needs to during boot.
Great. How about that opencores.org vga design, are you allergic to it? Is it overkill? Too much real estate? License problems? > My proposal is to point the A000:0000 memory block, which I think is > the "framebuffer" at some part of our graphics memory. Then a bit of > logic in our chip reads it and translates it to pixels so that the > video controller can handle it. A fair amount of memory bandwidth, > but only in that mode, and certainly fast enough to translate VGA in > real-time. Yep, that's the way to go all right. Now, if the host can map some of the video memory it can obviously map all of it. Surely this is how software rendering fallback is to be implemented? I suppose this raises the 1/W question again. Fortunately, we can teach Mesa about that if necessary. Regards, Daniel _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
